Update to 0.59 (latest commit)

- Change MASTER_SITES to GitHub, where upstream moved
- Apply some patches from the GitHub Pull Requests to make the port build
  (upstream seems to lost some files while migrating to GitHub) and patch
  some bugs
- Set licenses
- Upstream does not tag releases on GitHub, so it's all guessing what the
  real version is. 0.59 is what I found in the source code.
- Install examples; they are useful to test if the port is working
- Remove old patches, they are no longer needed
- Update WWW

4 files changed, 33 insertions, 30 deletions

diff --git a/science/crf++/Makefile b/science/crf++/Makefile
index f64ca10fce04..4ab25dadee10 100644
--- a/science/crf++/Makefile
+++ b/science/crf++/Makefile
@@ -2,18 +2,37 @@
 # $FreeBSD$
 
 PORTNAME=	crf++
-PORTVERSION=	0.54
-PORTREVISION=	1
+# 0.59 as defined in
+# https://github.com/taku910/crfpp/commit/d78526835e4cfcc4822156724edda8a3839327c2
+DISTVERSION=	0.59
 CATEGORIES=	science math
-MASTER_SITES=	SF/crfpp/crfpp/${PORTVERSION}
-DISTNAME=	CRF++-${PORTVERSION}
+PATCH_SITES=	https://patch-diff.githubusercontent.com/raw/taku910/crfpp/pull/:pr
+PATCHFILES=	15.patch:-p1:pr \
+		16.patch:-p1:pr
 
 MAINTAINER=	ports@FreeBSD.org
 COMMENT=	Yet Another CRF toolkit
 
-GNU_CONFIGURE=	yes
-INSTALL_TARGET=	install-strip
+LICENSE=	BSD3CLAUSE LGPL21+
+LICENSE_COMB=	dual
+LICENS_BSD3CLAUSE=	${WRKSRC}/BSD
+LICENS_LGPLv21+ =	${WRKSRC}/LGPL
+
 USES=		libtool
+USE_GITHUB=	yes
+GH_ACCOUNT=	taku910
+GH_PROJECT=	crfpp
+GH_TAGNAME=	1dc92a606f874a4fe52603803364cc1d90f952fb
 USE_LDCONFIG=	yes
+GNU_CONFIGURE=	yes
+INSTALL_TARGET=	install-strip
+
+PORTEXAMPLES=	JapaneseNE basenp chunking seg
+
+OPTIONS_DEFINE=	EXAMPLES
+
+post-install-EXAMPLES-on:
+	@${MKDIR} ${STAGEDIR}${EXAPLESDIR}
+	(cd ${WRKSRC}/example && ${COPYTREE_SHARE} . ${STAGEDIR}${EXAMPLESDIR})
 
 .include <bsd.port.mk>
diff --git a/science/crf++/distinfo b/science/crf++/distinfo
index 70e56cf6c118..af2b11f10ae2 100644
--- a/science/crf++/distinfo
+++ b/science/crf++/distinfo
@@ -1,2 +1,7 @@
-SHA256 (CRF++-0.54.tar.gz) = 8dab6e9ab71212b3d223588f6855384f2b29dada0a30d450266f8990ced63354
-SIZE (CRF++-0.54.tar.gz) = 809008
+TIMESTAMP = 1600806136
+SHA256 (taku910-crfpp-0.59-1dc92a606f874a4fe52603803364cc1d90f952fb_GH0.tar.gz) = 7f70e4113dd93caa34091e7f2df03ff75370934e3450c3a3b0d0a4afb11c8bd4
+SIZE (taku910-crfpp-0.59-1dc92a606f874a4fe52603803364cc1d90f952fb_GH0.tar.gz) = 864713
+SHA256 (15.patch) = eaf3408fcadf281eba05735af830acf82950a8da82e7cc565750a8e4d31d5480
+SIZE (15.patch) = 2376
+SHA256 (16.patch) = 5f27e76c8efac6c1fd4ac41c5925bda5b5c0063801a6a65fe311ff676e53cd1b
+SIZE (16.patch) = 98664
diff --git a/science/crf++/files/patch-feature__index.cpp b/science/crf++/files/patch-feature__index.cpp
deleted file mode 100644
index 042fafa6d42a..000000000000
--- a/science/crf++/files/patch-feature__index.cpp
+++ /dev/null
@@ -1,21 +0,0 @@
---- feature_index.cpp.orig	2018-09-01 10:26:57 UTC
-+++ feature_index.cpp
-@@ -34,8 +34,7 @@ int EncoderFeatureIndex::getID(const char *key) {
-   std::map <std::string, std::pair<int, unsigned int> >::iterator
-       it = dic_.find(key);
-   if (it == dic_.end()) {
--    dic_.insert(std::make_pair<std::string, std::pair<int, unsigned int> >
--                (key, std::make_pair<int, unsigned int>(maxid_, 1)));
-+    dic_.insert(std::make_pair(key, std::make_pair(maxid_, 1)));
-     int n = maxid_;
-     maxid_ += (key[0] == 'U' ? y_.size() : y_.size() * y_.size());
-     return n;
-@@ -171,7 +170,7 @@ void EncoderFeatureIndex::shrink(size_t freq) {
-     const std::string &key = it->first;
- 
-     if (it->second.second >= freq) {
--      old2new.insert(std::make_pair<int, int>(it->second.first, new_maxid));
-+      old2new.insert(std::make_pair(it->second.first, new_maxid));
-       it->second.first = new_maxid;
-       new_maxid += (key[0] == 'U' ? y_.size() : y_.size() * y_.size());
-       ++it;
diff --git a/science/crf++/pkg-descr b/science/crf++/pkg-descr
index 74979ee280eb..dfc641cdb775 100644
--- a/science/crf++/pkg-descr
+++ b/science/crf++/pkg-descr
@@ -5,4 +5,4 @@ designed for generic purpose and will be applied to
 a variety of NLP tasks, such as Named Entity Recognition,
 Information Extraction and Text Chunking.
 
-WWW: http://chasen.org/~taku/software/CRF++/
+WWW: https://taku910.github.io/crfpp/